Re: [sv-cc] Polls on extern/export and representation of SV


Subject: Re: [sv-cc] Polls on extern/export and representation of SV
From: Andrzej Litwiniuk (Andrzej.Litwiniuk@synopsys.com)
Date: Fri Mar 07 2003 - 16:50:51 PST


> > > 1. Poll on acceptance of Joao's modified extern/export proposal.
> > > http://www.eda.org/sv-cc/hm/0920.html

ABSTAIN

        extern part
        I don't mind the syntax (personal dislike for "DPI" is a non-issue).
        I like the ability to have externs inside any scope, what is great for
        libraries.
        I like very much the ability to use a local declaration for specifying
        default values, individually for each scope.
        I don't like the obligatory names of formal arguments and I see
        as a serious inconvenience the elimination of multiple declarations
        in the topmost level. (I'm with you, Kevin!)

        export part
        My understanding (see draft of LRM), apparently wrong, was that
        an export declaration would provide a full hierarchical name to
        the SV function instance. (same opinion as Francoise's)
        I can see both the need for and the power of context-sensitive calls,
        but also can see implementation issues (the cost of veryfing the
        correctness of a SV call from C and of locating SV code).

        I don't like the fact that the export declarations will be scattered
        over the design and possibly buried deep inside.
        There is no way to have all exports grouped together in some
        well visible/exposed place, in $root or at the top level of a library
        module (if module exports something defined deeper/lower).

> Francoise:
> [...] I am still leaning towards preferring an export
> declaration which [...] would provide a full hierarchical
> name to the SV function instance.

> > > 2. Poll on agreement on SV-CC's opinion on the SV LRM language
> > > issue. Andrzej's comment on this is at:
> > > http://www.eda.org/sv-cc/hm/0908.html

ABSTAIN

        I still don't know the answer to my dillema concerning
        the representation of unpacked arrays of packed types.

> Francoise:
> However Andrzej's proposal is not assertive on unpacked arrays
         of packed types.

        Yes!

Regards,
Andrzej



This archive was generated by hypermail 2b28 : Fri Mar 07 2003 - 16:51:39 PST