Re: FlexLocks

Robert Haas wrote:
> On Wed, Nov 16, 2011 at 12:25 PM, Kevin Grittner
> <Kevin(dot)Grittner(at)wicourts(dot)gov> wrote:
> >> We could alternatively change one or the other of them to be a
> >> struct with one member, but I think the cure might be worse than
> >> the disease. ?By my count, we are talking about saving perhaps as
> >> many as 34 lines of code changes here, and that's only if
> >> complicating the type handling doesn't require any changes to
> >> places that are untouched at present, which I suspect it would.
> >
> > So I stepped through all the changes of this type, and I notice that
> > most of them are in areas where we've talked about likely benefits
> > of creating new FlexLock variants instead of staying with LWLocks;
> > if any of that is done (as seems likely), it further reduces the
> > impact from 34 lines. ?If we take care of LWLockHeldByMe() as you
> > describe, I'll concede the FlexLockId changes.
>
> Updated patches attached.

It would be helpful if the patch included some text about how flexilocks
are different from ordinary lwlocks.

--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com

+ It's impossible for everything to be true. +