RE: [sv-ac] IEEE 1800 SV-AC : minutes of meeting on 10/10 2006

From: Miller Hillel-R53776 <r53776_at_.....>
Date: Wed Oct 11 2006 - 00:31:02 PDT
Ed,

I think the intention is not to provide formal/actual arguments of a
sequence when specifying a sequence as formal type. As doron an dmitry
said the formal argument should not include the signiture. The sequence
type can be assigned with any legal sequence instance. So the example
should reduce down to this.

sequence s1(int flv);
  (exp1, flv = exp2);
endsequence

sequence s2(sequence s); // formal arg on a formal arg?
  int lv;
  s ##1 (lv == exp3); // lv as arg??
endsequence
property p;
  s2(s1(lv));
endproperty 

My concern in this example is that the user whom is using the sequence
s2 needs to be aware of s2's local variables .  If we do:

property p;
  s2(s1(lv_arg));
endproperty 

Will this result in a undeclared variable error?

If this is the case then this use model is problematic from a language
perspective. 
I think it is generally expected in a modular programming framework to
have a well defined interface as part of the code, that is, in this
example one would need to specify how to use sequence s2. Not having the
name of the local variable specified in the header of the sequence does
not pack the sequence s2 well for external usage.

Thanks
Hillel

-----Original Message-----
From: owner-sv-ac@eda.org [mailto:owner-sv-ac@eda.org] On Behalf Of
Eduard Cerny
Sent: Tuesday, October 10, 2006 7:44 PM
To: sv-ac@eda.org
Subject: [sv-ac] IEEE 1800 SV-AC : minutes of meeting on 10/10 2006

Please find attached the minutes of the SV-AC meeting earlier today. Let
me know if you see any discrepancies.

Best regards,
ed
Received on Wed Oct 11 00:31:09 2006

This archive was generated by hypermail 2.1.8 : Wed Oct 11 2006 - 00:32:41 PDT