RE: More analysis of IR2099

Chuck,

The analysis looks very good. I like the ideas embodied in principles 1 and
2.

One question that I alluded to last telecon is that, while the principles
avoid the issue of legality of an alias in the same declarative region as
the original declaration, it could introduce ambiguities. To use a variant
of case 1 as an illustration:

package p_test is
    type my_logic is ( a, b, c, d);
    alias alt_logic is my_logic;
end package p_test;

(I've changed the enumeration literals to avoid clouding the issue with bit
and character literals.) This declares, among other things:

  "="[my_logic, my_logic return boolean]

and

  "="[alt_logic, alt_logic return boolean]

The question is whether an expression like a = b is now ambiguous. I think
the fact that a and b are both of the same type and that both "=" operators
have the same profile and denote the same function should mean that the
expression is not ambiguous. But to make this 100% clear, we may need to add
wording to the rules for overload resolution and for profile comparison to
say that an alias denotes the same thing as the aliased name, and that this
fact can be used to resolve ambiguity in favour of the aliased name. (I see
now you've left that an open issue for 10.5 in the recommendations.)

A minor point on the recommentations. In the definition of an identity
alias, we should say the alias designator is the same as the simple name,
character literal, or operator symbol of the denoted name. (See,
incidentally, the IR I submitted today. I think that might also address your
question about enumerations.)

Looks like we're closing in on what's turned out to be a thorny issue.
That's what makes it fun to be part of ISAC! ;-)

Cheers,

PA

--
Dr. Peter J. Ashenden         peter@ashenden.com.au
Ashenden Designs Pty. Ltd.    www.ashenden.com.au
PO Box 640                    VoIP: sip://0871270078@sip.internode.on.net
Stirling, SA 5152             Phone (mobile):  +61 414 70 9106
Australia


> -----Original Message-----
> From: owner-isac@server.eda.org 
> [mailto:owner-isac@server.eda.org] On Behalf Of Chuck Swart
> Sent: Friday, 8 December 2006 7:23 AM
> To: isac@server.eda.org
> Subject: More analysis of IR2099
> 
> 
> This isn't complete, but here's what I have so far:
> 
> 
>