RE: [sv-bc] FW: Can a keyword be used as identifier if context is clear?

From: Clifford E. Cummings <cliffc_at_.....>
Date: Fri May 12 2006 - 10:16:41 PDT
Ouch!!

The reason we have 1364-2001-noconfig is because one vendor 
implemented the 2001 standard and allowed all of the config keywords 
to be used as identifiers in a non-config context. I don't remember 
if this was a misinterpretation of the original Verilog-2001 standard 
or if the vendor did no get around to implementing configs until 
after other 2001 features had been implemented and by then users were 
using 2001 features with config and cell keywords in existing 
designs. The vendor prevailed on the committee to at least allow a 
switch that would permit those existing customers to continue to use 
Verilog-2001 features without the need to remove the offending 
config-subset of keywords. It was a rather heated debate but in the 
end the committee voted to bail-out the vendor through the use of 
this switch-variant.

When I put together the original Verilog-2001 annex B (keywords) 
section, I had separated the config keywords into a separate list of 
keywords so that the config keywords could be used in a design, but 
the committee very deliberately made me change the keyword lists into 
a single keyword list because nobody wanted to have context-sensitive 
keywords. I provided clear evidence of this intent a year ago when we 
dated this topic, yet again.

Now it is beginning to look like the bail-out is turning into a 
full-blown wart. Do we really want to be in the business of defining 
subset compiler directives to choose which key words are important to 
individual users? Do we really want to require that users be 
cognizant of keyword sub-lists that can be used with specific 
compiler directives??

Really folks! Let's not get into the business of 
compiler-directive-directed keyword sub-lists! I think the current 
list of keyword compatibilities are more than sufficient (I still get 
a giggle every time I see the 1364-2001-noconfig switch!)

Regards - Cliff

At 07:25 AM 5/12/2006, Bresticker, Shalom wrote:
>So why should 1364-2001 be different?
>
>Shalom
>
> > -----Original Message-----
> > From: owner-sv-bc@server.eda.org [mailto:owner-sv-bc@server.eda.org]
>On
> > Behalf Of Stuart Sutherland
> > Sent: Friday, May 12, 2006 5:16 PM
> > To: sv-bc@server.eda.org
> > Subject: RE: [sv-bc] FW: Can a keyword be used as identifier if
>context
> > is clear?
> >
> > Brad,
> >
> > The 1364-2005 standard already has the following:
> >
> > version_specifier ::=
> > | 1364-1995
> > | 1364-2001
> > | 1364-2001-noconfig
> > | 1364-2005
> >
> > The SV standard adds 1800-2005.  No one suggested an "1800-2005-
> > noconfig"
> > switch, and I didn't think of it, either, when I wrote the proposal.
>In
> > my
> > opinion, any tool that supports the SV standard should reserve the
> > entire SV
> > keyword list, with no exceptions.
> >
> > ~~~~~~~~~~~~~~~~~~~~~~~~~
> > Stuart Sutherland
> > stuart@sutherland-hdl.com
> > +1-503-692-0898

----------------------------------------------------
Cliff Cummings - Sunburst Design, Inc.
14314 SW Allen Blvd., PMB 501, Beaverton, OR 97005
Phone: 503-641-8446 / FAX: 503-641-8486
cliffc@sunburst-design.com / www.sunburst-design.com
Expert Verilog, SystemVerilog, Synthesis and Verification Training
Received on Fri May 12 10:16:36 2006

This archive was generated by hypermail 2.1.8 : Fri May 12 2006 - 10:16:41 PDT