On Tue, 2006-03-21 at 18:13 +0100, Duncan Sands wrote: > Is memory use a problem here? VR_VARYING has the advantage of using > a bit less memory. On the other hand, I guess you could introduce the > convention that VR_RANGE with null min/mae means to use TYPE_MIN/ > TYPE_MAX, or something along those lines. Well, more than anything, I just don't think we really thought about the issue of what the best representation would be in this kind of case.
If I was to lean any particular direction I would tend to lean towards exposing the range and removing VR_VARYING. The net result would be good consistency in the transformations within VRP. ie, we would not need to special case stuff like VR_VARYING | ~[0, 0] -> ~[0, 0] SImilarly for typecasts and a variety of other stuff. jeff > > Ciao, > > D.