>Yes, I agree that should be done. But now that SV has a more general
>casting mechanism, its seems more natural to use the casting syntax to
>change the signing rather than a system function.
But the semantics of casting are only fully defined for casting to
full types (that business about being equivalent to assigning to an
intermediate variable of the type). The semantics of casting to partial
types (just width or signedness) are not fully specified.
Such partial typecasts may be desirable, and rules for them could certainly
be specified. But at present they aren't.
Steven Sharp
sharp@cadence.com
Received on Wed Jun 23 11:27:34 2004
This archive was generated by hypermail 2.1.8 : Wed Jun 23 2004 - 11:27:47 PDT