>From: "Feldman, Yulik" <yulik.feldman@intel.com> >I was thinking about using the definition of matching types for this >purpose. It is already defined and may be OK to use. It does look reasonable for this. > The algorithm is then: > >1. If all types match, return one of them. >2. If they don't match, and they are not integral, it is an error. >3. If they don't match, and they are integral, return the normalized >type. > >Isn't this simple enough? Your rule includes matching the type of the context for integral types. So the next thing you would have to do is define what the type of the context is. The existing LRM only defines the size and signedness of the context. After that, there will be other issues that require more special exceptions to your rules. For example, your rules produce the wrong result for 2-state types. If you have int i, j; type(cond ? i : j) your rules would say that the type is int. But the result is definitely a 4-state vector, so it cannot be type int. Steven Sharp sharp@cadence.com -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.Received on Wed Oct 24 12:31:05 2007
This archive was generated by hypermail 2.1.8 : Wed Oct 24 2007 - 12:31:34 PDT