| Number | 155
|
| Category | errata
|
| Synopsis | A.1.3, module_declaration ambiguous
|
| State | lrmdraft
|
| Class | errata-simple
|
| Arrival-Date | Oct 09 2002
|
| Originator | "Brad Pierce" <Brad.Pierce@synopsys.com>
|
| Release | 2001b: 2.8.2 (Syntax 2-4), 12.1 (Syntax 12-1), A.1.3
|
| Environment |
|
| Description |
In A.1.3, module_declaration is ambiguous. The fix is to remove the [] from around list_of_port_declarations in the second option -- [list_of_port_declarations] --> list_of_port_declarations To see the ambiguity, consider a case where there is neither a list_of_ports nor a list_of_port_declarations and, moreover, after the semicolon there are no port_declarations. Then it's ambiguous whether there's a {module_item} or a {non_port_module_item} and which of the options was used to generate the module_declaration. The proposed fix does more than just remove ambiguity. It also more accurately reflects the actual intent of the {non_port_module_item} restriction, which should be imposed only in the case when there is a list_of_port_declarations. Otherwise, a full {module_item} is allowed. -- Brad |
| Fix |
Section A.1.3 Passed 11/18/02 by ETF: Remove the square brackets around the list_of_ports element in the old style module_declaration production. REPLACE: module_declaration ::= { attribute_instance } module_keyword module_identifier [ module_parameter_port_list ] [ list_of_ports ] ; { module_item } endmodule | { attribute_instance } module_keyword module_identifier [ module_parameter_port_list ] [ list_of_port_declarations ] ; { non_port_module_item } endmodule WITH: module_declaration ::= { attribute_instance } module_keyword module_identifier [ module_parameter_port_list ] list_of_ports ; { module_item } endmodule | { attribute_instance } module_keyword module_identifier [ module_parameter_port_list ] [ list_of_port_declarations ] ; { non_port_module_item } endmodule |
| Audit-Trail |
From: Shalom Bresticker <Shalom.Bresticker@motorola.com> To: Brad Pierce <Brad.Pierce@synopsys.com> Cc: etf-bugs@boyd.com Subject: Re: errata/155: A.1.3, module_declaration ambiguous Date: Thu, 10 Oct 2002 09:33:13 +0200 >Category: errata >Confidential: no >Originator: Shalom Bresticker <Shalom.Bresticker@motorola.com> >Release: 2001b >Class: TBD >Description: agreed This affects also Syntax 2-4 in 2.8.2 and Syntax 12-1 in 12.1. From: "Brad Pierce" <Brad.Pierce@synopsys.com> To: <etf-bugs@boyd.com> Cc: Subject: Re: errata/155: A.1.3, module_declaration ambiguous Date: Mon, 11 Nov 2002 15:34:34 -0800 >Category: errata >Confidential: no >Originator: "Brad Pierce" <Brad.Pierce@synopsys.com> >Release: 2001b >Class: TBD >Description: Dan Jacobi points out that the correct fix is to remove the [] from around the list_of_ports instead, because if there is neither a list of ports nor a list of port declarations, then there should be no port declarations in the body. If a missing list is treated as an empty list of port declarations, this will enforce that only non_port_module_items are allowed. -- Brad From: Steven Sharp <sharp@cadence.com> To: etf-bugs@boyd.com, Brad.Pierce@synopsys.com Cc: Subject: Re: errata/155: A.1.3, module_declaration ambiguous Date: Mon, 11 Nov 2002 18:46:16 -0500 (EST) >Category: errata >Confidential: no >Originator: Steven Sharp <sharp@cadence.com> >Release: 2001b >Class: TBD >Description: > Dan Jacobi points out that the correct > fix is to remove the [] from around the > list_of_ports instead, because if there > is neither a list of ports nor a list of > port declarations, then there should be no > port declarations in the body. If a missing > list is treated as an empty list of port > declarations, this will enforce that only > non_port_module_items are allowed. The existing semantic rule that port declarations in the body must match the list_of_ports will enforce that anyway. As with any proper definition, the degenerate case works properly in either case. Steven Sharp sharp@cadence.com From: "Brad Pierce" <Brad.Pierce@synopsys.com> To: <etf-bugs@boyd.com> Cc: Subject: Re: errata/155: A.1.3, module_declaration ambiguous Date: Mon, 11 Nov 2002 16:09:15 -0800 >Category: errata >Confidential: no >Originator: "Brad Pierce" <Brad.Pierce@synopsys.com> >Release: 2001b >Class: TBD >Description: Whenever there's an opportunity to enforce semantic rules with syntactic rules, I'd like to exploit it. In the case of this erratum, a change is needed to remove ambiguity from the grammar. The change of removing the [] from list_of_ports expresses the intended semantics better than does the change of removing the [] from list_of_port_declarations. -- Brad From: Steven Sharp <sharp@cadence.com> To: etf-bugs@boyd.com, Brad.Pierce@synopsys.com Cc: Subject: Re: errata/155: A.1.3, module_declaration ambiguous Date: Mon, 11 Nov 2002 19:50:14 -0500 (EST) >Category: errata >Confidential: no >Originator: Steven Sharp <sharp@cadence.com> >Release: 2001b >Class: TBD >Description: > The change of > removing the [] from list_of_ports expresses the > intended semantics better than does the change of > removing the [] from list_of_port_declarations. Actually, I disagree. The 1995 standard allowed an empty port list. That was clearly an empty list_of_ports. It is easier to see what the 2001 standard added if you have one rule for the old syntax and another rule for the new. And if someone is linting for features that are not backward compatible with 1995, they had better not treat the empty port list as a 2001 list_of_port_declarations. Removing the [] from list_of_port_declarations expresses the change in syntax better. And since they both have identical semantics, neither way expresses the semantics any better. Steven Sharp sharp@cadence.com From: Shalom Bresticker <Shalom.Bresticker@motorola.com> To: Steven Sharp <sharp@cadence.com> Cc: etf-bugs@boyd.com Subject: Re: errata/155: A.1.3, module_declaration ambiguous Date: Sun, 17 Nov 2002 11:50:50 +0200 >Category: errata >Confidential: no >Originator: Shalom Bresticker <Shalom.Bresticker@motorola.com> >Release: 2001b >Class: TBD >Description: I'll go with Brad on this one. He is correct about tightening up the BNF and it comes for free. The 2001 BNF is so different from the 1995 BNF in so many places that a simplistic comparison using 'diff' on the BNF is not realistic. In fact, as I recently wrote, we have a number of change proposals on the table that, if accepted, will cause that to occur even from 2001 to 2004. But in this case, I won't make a big thing of it because Steven is also correct about the semantics. So what is most important is that both ways of fixing the ambiguity work. Shalom Steven Sharp wrote: > > The change of > > removing the [] from list_of_ports expresses the > > intended semantics better than does the change of > > removing the [] from list_of_port_declarations. > > Actually, I disagree. The 1995 standard allowed an empty > port list. That was clearly an empty list_of_ports. It > is easier to see what the 2001 standard added if you have > one rule for the old syntax and another rule for the new. > And if someone is linting for features that are not > backward compatible with 1995, they had better not treat > the empty port list as a 2001 list_of_port_declarations. > > Removing the [] from list_of_port_declarations expresses > the change in syntax better. And since they both have > identical semantics, neither way expresses the semantics > any better. -- Shalom Bresticker Shalom.Bresticker@motorola.com Design & Reuse Methodology Tel: +972 9 9522268 Motorola Semiconductor Israel, Ltd. Fax: +972 9 9522890 POB 2208, Herzlia 46120, ISRAEL Cell: +972 50 441478 "The devil is in the details." |
| Unformatted |
|
Hosted by Boyd Technology