Re: Reducing some DDL Locks to ShareLock

On Tue, 2008-11-18 at 15:04 -0500, Tom Lane wrote:
> Simon Riggs <simon(at)2ndQuadrant(dot)com> writes:
> > heap_inplace_fetch takes as input "tuple" which is a palloc'd tuple,
> > extracts from it the tid of the tuple, reads the buffer, locks it, then
> > releases the original tuple. It then returns a copy of the on-block
> > tuple. So all other code the same as before when we were working on a
> > copy produced from the syscache.
>
> > Is that roughly what you intended?
>
> I'd suggest making it take a TID rather than presuming where the caller
> is going to get the TID from. Otherwise, +1.

Just to mention I haven't forgotten about this. I wrote a patch on 20/11
and was debugging it when I fell ill. I've updated the patch now to CVS
HEAD and will retest it in next few days.

--
Simon Riggs www.2ndQuadrant.com
PostgreSQL Training, Services and Support