Amit Kapila <amit.kapil...@gmail.com> writes:
> Another point which I am wondering is why can't we use the existing
> REWIND flag to solve the current issue, basically if we have access to
> that information in nodeLimit.c (ExecLimit), then can't we just pass
> down that to ExecShutdownNode?
The existing REWIND flag tells subnodes whether they should *optimize*
for getting rewound or not. I don't recall right now (well past
midnight) why that seemed like a useful definition, but if you grep for
places that are paying attention to that flag, I'm sure you'll find out.
We probably don't want to give up that distinction --- if it had been
equally good to define the flag as a hard yes-or-no, I'm sure we would
have taken that definition, because it's simpler.
regards, tom lane
