RE: [sv-ac] review of 1901 (constant expression in ##)

Yes sorry you are absolutely correct,
 
I was only looking in the constant_expression or
const_or_range_expression and missed that part.
 
So the proposal is fine.
 
 
Thanks for your quick perception,
Yaniv
 


________________________________

From: Bresticker, Shalom [mailto:shalom.bresticker@intel.com] 
Sent: Wednesday, January 16, 2008 18:09
To: Fais Yaniv; sv-ac@eda.org
Subject: RE: [sv-ac] review of 1901 (constant expression in ##)


Yaniv,
 
I don't understand (maybe because I don't pay continuous attention).
 
Currently the syntax of cycle_delay_range is
 
cycle_delay_range ::=

## integral_number

| ## identifier

| ## ( constant_expression )

| ## [ cycle_delay_const_range_expression ]

which allows you to specify any identifier, including of a variable.
 
I seem to remember that this was an issue with passing a
cycle_delay_range to an assertion construct via its argument list rather
than as a parameter.
 
Regards,
Shalom


________________________________

	From: owner-sv-ac@server.eda.org
[mailto:owner-sv-ac@server.eda.org] On Behalf Of Fais Yaniv
	Sent: Wednesday, January 16, 2008 5:51 PM
	To: Eduard Cerny; sv-ac@server.eda.org
	Subject: [sv-ac] review of 1901 (constant expression in ##)
	
	
	Hi,
	 
	I have reviewed 1901 proposal and actually found it redundant,
	 
	It is claimed  that "Cycle delay for ## concatenation allows
identifier to specify the delay w/o retsricting to constant epxr"
	 
	The addition this proposal makes is:
	 
	- When an identifier is specified as in ##N, the identifier must
be a constant primary, i.e., an

	elaboration-time constant.

	 

	currently the BNF for constant_expression is:
	 
	constant_expression ::=

	constant_primary

	| unary_operator { attribute_instance } constant_primary

	| constant_expression binary_operator { attribute_instance }
constant_expression

	| constant_expression ? { attribute_instance }
constant_expression : constant_expression

	 

	constant_primary ::=

	primary_literal

	| ps_parameter_identifier constant_select

	...

	 

	(primary literal is more or less a number)

	 

	so when an identifier is used it must already be part of
constant_primary (parameter,genvar and etc.).

	 

	 

	It may still be a good idea to add this sentence to make it
clearer regardless of the BNF.

	 

	Regards,

	Yaniv

	 


	-- 
	This message has been scanned for viruses and 
	dangerous content by MailScanner <http://www.mailscanner.info/>
, and is 
	believed to be clean. 


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