> AFAIK hardly anyone actually uses CURRENT, and I've been thinking of
> proposing that we eliminate it to make the world safe for constant-
> folding timestamp operations. (Thomas, any comments here?)
Well, it is a feature from "the old days". Pretty neat one at that, and
is an example of a useful feature not found in other DBs or in
standards, but which might show up someday because they are useful.
Throwing those things away one at a time will end us up at the lowest
common denominator, eventually :(
Another way of looking at the problem is to ask how we could retain this
feature in the face of the other optimization "desirements". istm that
types which have multiple behaviors could be queried for the behavior of
a particular example by the optimizer. For most types, a "query" would
not be necessary (so there is minimal overhead), but for this case a
function could return the property of an example as either cachable or
not.
Perhaps a true "serial type" would need similar behaviors, as might
other future types.
- Thomas