Reply To: Negative weights and index range in DistInt(weights)?

[Jim Lewis]

Hi Andy,

Issue 3:  return a value in the range of the weight vector

This would not impact any of the use models I have looked at as I typically have use positional aggregates.  If I remember correctly, named aggregates of arrays always follow the direction of the index base type (natural), and hence, default to a range of “to”.  

It would mean rewriting the input processing algorithm to handle either a “to” or a “downto” range, but that should be do able.  

If a tool bug were to mess up the range of an aggregate literal, it would be interesting debugging.  That should not drive the implementation though.

Also see DistValInt.  This addresses the general case of what you are thinking of above.  Name is different due to lack of introspection of composites.   

Best,

Jim