[sv-cc] Mantis items 1503 and 2237 (VPI parts)

Bassam,

I was looking over the proposal for Mantis 2237, and it reminded me of 
the concern I had about the proposal for Mantis 1503, since 2237 seems 
to assume 1503.

In both proposals the property formal decl and sequence formal decl 
have a vpiExpr relation to various kinds of objects.  We did this in an 
attempt to emulate what happens for io decls in tasks and functions.  I 
don't think this works, though.

In practice, because a property formal decl or sequence formal decl can 
be untyped, what it maps to "internally" also may not have a type.  The 
same property formal decl, when compared to a matching property inst, 
could wind up being connected to a net of any type, a variable of any 
type, a sequence, etc.  So given that we're working with a single 
property decl and potentially multiple property insts, we cannot have a 
single vpiExpr relation from the formal decl and have it mean anything.

On the other hand, a property formal decl or sequence formal decl may 
have a data type specified as part of the property decl, independently 
of the property insts.  So it would make sense to have a vpiTypespec 
relation from a property formal decl or sequence formal decl, but not a 
vpiExpr.  If the formal decl is untyped, the vpiTypespec would of 
course return NULL.

May I recommend making that change to both proposals?

Regards,
Jim Vellenga

--------------------------------------------------------- 
James H. Vellenga                            978-262-6381 
Software Architect                     (FAX) 978-262-6636 
Cadence Design Systems, Inc.         vellenga@cadence.com 
270 Billerica Rd
Chelmsford, MA 01824-4179
"We all work with partial information." 
---------------------------------------------------------- 

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.