I said:
> In this particular case, since the qual doesn't actually touch the
> type-converted column, it would have been safe to push down, but the
> planner doesn't make this test on a per-qual basis: if there are type
> conversions anywhere in the UNION it just punts.
CVS tip has now had its consciousness raised on this point ;-). Turns
out that it's not significantly more expensive to apply the test to
individual vars instead of having a blanket check on the whole query,
so I fixed it.
Jonathan will still need to put the explicit casts into his view for
use with 7.3.*, though. He probably should do that anyway, because
even in 7.4 a restriction on the "type" column would not get pushed
down as desired if the literals are UNKNOWN type.
regards, tom lane