[sv-bc] Re: [sv-ec] Query for pattern matching

Moumita,

I don't think there is any issue with allowing hierarchical
references within the cond_pattern expression.  There isn't
any fundamental circularity that I see -- there might be
artificial circularities in a simplistic elaboration approach
but I don't see anything fundamental.  The types of the
newly scope identifiers progress linearly through the
pattern nestings and cannot be referenced from outside the
scope of that condition.

Can you give a better explanation or an example of where you
believe there is a circularity?

Gord.


Moumita wrote:
> Hello,
> 
> We know that  hierarchical resolution are aspects of elaboration activity.
> According to the BNF of pattern matching mentioned in IEEE Std 
> 1800-2005, IEEE Standard for SystemVerilog:-
> cond_pattern ::= expression matches pattern
> pattern ::= . variable_identifier
> 
> variable_identifier is created  using the type of the expression 
> mentioned in cond_pattern.
> and LRM does not restrict  for the expression so we can  assume that  
> hierarchical reference  can be used as expression.
> 
> But if   hierarchical reference is used as expression in cond_pattern 
> then variable_identifier is created
> after resolving that  hierarchical reference at the time of elaboration 
> and all the semantic checks depending upon that variable_identifier will 
> be done there after.
> That can raise cyclic dependency in some situations.
> 
> So my query is - whether hierarchical reference is allowed as 
> expression  in cond_pattern  or not?
> and for the following testcase ,where  hierarchical reference is used as 
> an expression
> in cond_pattern is valid or not?
> 
> module mod;
>    bit i1[8:0];
>    bit k1[8:0];
>    bit j1[8:0];
>    initial begin
>        k1 = mod.i1 matches .n? n : j1;
>    end
> ...
> 
> Thanks and regards,
> 
> Moumita
> 
> 
> 

-- 
--------------------------------------------------------------------
Gordon Vreugdenhil                                503-685-0808
Model Technology (Mentor Graphics)                gordonv@model.com


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