Comment 8 for bug 1253163

IMO, we should go ahead and mark it STABLE and STRICT now. If we have cause to use the function more than once in a query, there's no need to recalculate the output for the same parameters. (In fact, we wouldn't want that anyway.) Since we're knocking around this function's uses in this bug, doing it here is probably less surprising (or, at least more context-y) than in a follow-on script doing general work.