On Sat, Feb 28, 2015 at 2:45 PM, Stephen Frost <sfrost@snowman.net> wrote:
> I'm trying to wrap my head around the reasoning for this also and not
> sure I'm following. In general, I don't think we protect all that hard
> against functions being called with tokens that aren't allowed by the
> parse.
Check.
> So, basically, this feels like it's not really the right place
> for these checks and if there is an existing problem then it's probably
> with the grammar... Does that make sense?
As long as there is no more inconsistency between the parser, that
sometimes does not set VACOPT_FREEZE, and those assertions, that do
not use the freeze_* parameters of VacuumStmt, I think that it will be
fine.
[nitpicking]We could improve things on both sides, aka change gram.y
to set VACOPT_FREEZE correctly, and add some assertions with the
params freeze_* at the beginning of vacuum().[/]
--
Michael