Subject: A.1.4, A.7.3: Should range_expression be constant_range_expression?
From: Shalom.Bresticker@motorola.com
Date: Mon Dec 30 2002 - 23:59:23 PST
>Category: errata
>Synopsis: A.1.4, A.7.3: Should range_expression be constant_range_expression?
>Confidential: no
>Severity: serious
>Priority: medium
>Class: errata-discuss
>Submitter-Id: etf
>Originator: Shalom.Bresticker@motorola.com
>Release: 2001b: 12.1 (Syntax 12-1), 12.3.1 (Syntax 12-5), 14.2.2 (Syntax 14-3), A.1.4, A.7.3
>Environment:
#123
SV-BC19-45,47
>Description:
There are a few BNFs which offer as indexing alternatives
either [ constant_expression ] or [ range_expression ].
Specifically,
port_reference ::= (A.1.4, Syntax 12-1, 12-5)
port_identifier
| port_identifier [ constant_expression ]
| port_identifier [ range_expression ]
specify_input_terminal_descriptor ::= (A.7.3, Syntax 14-3)
input_identifier
| input_identifier [ constant_expression ]
| input_identifier [ range_expression ]
specify_output_terminal_descriptor ::= (A.7.3, Syntax 14-3)
output_identifier
| output_identifier [ constant_expression ]
| output_identifier [ range_expression ]
My immediate question is:
range_expression includes expression,
which is not restricted to be a constant.
constant_expression is therefore a proper subset of
range_expression.
Either range_expression should have been
constant_range_expression or
constant_expression should have been expression.
A related issue that will soon be filed comes from
SV-BC19-45 and 19-47 which points out that
constant_range_expression includes constant_expression
and range_expression includes expression,
so why do non-terminals such as net_lvalue and primary and
variable_lvalue and the 3 above,
which use constant_range_expression or range_expression,
need constant_expression or expression as well?
This also seems to be related to issue #123.
>How-To-Repeat:
>Fix:
>Unformatted:
This archive was generated by hypermail 2b28 : Mon Dec 30 2002 - 23:59:44 PST