Re: [COMMITTERS] pgsql: Use a bitmask to represent role attributes

* Alvaro Herrera (alvherre(at)2ndquadrant(dot)com) wrote:
> Bruce Momjian wrote:
>
> > I am with Tom on this --- there is more wasted space in the 'name'
> > column pg_authid.rolname than by shoving 40 boolean values into a
> > bitmap. Adding the complexity of a bitmap doesn't make sense here. I
> > also apologize for the late feedback.
>
> Okay, it seems we have a majority that does not want this patch -- at
> least Tom, Robert and Bruce plus-one'd the reversion. I am going to
> revert it, mainly because I don't want to be on the hook for fixing it
> later on. I'm also going to mark it Rejected in commitfest.

Well, I'd be happy for fixing it, but that's not what is at issue here.
If it was only about getting someone to maintain it, I don't think
there'd be a discussion.

> Adam and Stephen can rework as they see fit, according to whatever
> acceptable design is found.

In the end, this is simpler, and the patch to add the other role
attributes which we were looking to add is probably closer to being done
with this outcome anyway, since it doesn't need to be rebased on top of
the bitmap work.

Thanks,

Stephen