Re: [sv-ac] #1532

It doesn't work that way.  The semantic prefixes in the BNF are
normative.  For example, if 'foo' is declared as a function, it could
not be parsed as a task_identifier.
 
I'd also like to add that the grammar for event_control is not
ambiguous.  The LRM is not required to provide a BNF that can simply be
copied into Yacc.  An LALR(1) parser generator is going to experience
many conflicts if its input specification follows the BNF too closely.
To get around the limitations of LALR(1) it's often necessary to accept
a superset of the legal syntax, then whittle away illegal syntax with
semantic checks.
 
The Mantis item is not strictly speaking an erratum.  It is a request to
make parsing easier. 
 
-- Brad



________________________________

From: Eduard Cerny 
Sent: Monday, July 31, 2006 12:43 PM
To: Brad Pierce; sv-ac@eda-stds.org
Subject: RE: [sv-ac] #1532


Not directly, but the LRM says:
 
hierarchical_event_identifier ::= hierarchical_identifier

hierarchical_identifier ::= [ $root . ] { identifier constant_bit_select
. } identifier

 

and a sequence name is an identifier.

 

ed


________________________________

	From: owner-sv-ac@eda-stds.org [mailto:owner-sv-ac@eda-stds.org]
On Behalf Of Brad Pierce
	Sent: Monday, July 31, 2006 3:39 PM
	To: sv-ac@eda-stds.org
	Subject: Re: [sv-ac] #1532
	
	
	A sequence is a kind of event?
	 
	-- Brad

________________________________

	From: owner-sv-ac@eda-stds.org [mailto:owner-sv-ac@eda-stds.org]
On Behalf Of Kulshrestha, Manisha
	Sent: Monday, July 31, 2006 12:35 PM
	To: Brad Pierce; sv-ac@eda-stds.org
	Subject: RE: [sv-ac] #1532
	
	
	Hi,
	 
	This proposal does not break backward compatibility as
@hierarchical_event_identifier takes care of the case 
	a = @s c;
	 
	What this proposal is saying is that if there is a parameterized
sequence instance, it has to be inside '(..)' so that it does not create
any conflicts.
	 
	Thanks.
	Manisha

________________________________

	From: owner-sv-ac@server.eda-stds.org
[mailto:owner-sv-ac@server.eda-stds.org] On Behalf Of Brad Pierce
	Sent: Monday, July 31, 2006 12:13 PM
	To: sv-ac@server.eda-stds.org
	Subject: Re: [sv-ac] #1532
	
	
	The proposed BNF change unnecessarily breaks backward
compatibility.  One would no longer be able to write
	 
	     a = @s c;
	 
	for a sequence 's'.  Yet, by replacing 'sequence_instance' with
'ps_sequence_identifier' in 'event_control', as suggested in Mantis on
7-11-2006, you could preserve this part of backward compatibility and
still address the easier-to-Yacc desire of the erratum.
	 
	
http://www.eda-stds.org/svdb/bug_view_page.php?bug_id=0001532
	 
	-- Brad

________________________________

	From: owner-sv-ac@eda-stds.org [mailto:owner-sv-ac@eda-stds.org]
On Behalf Of Kulshrestha, Manisha
	Sent: Monday, July 31, 2006 11:19 AM
	To: sv-ac@eda-stds.org
	Subject: [sv-ac] #1532
	
	
	Hi,
	 
	I have uploaded a proposal for #1532. 
	 
	Thanks.
	Manisha