Subject: RE: [sv-cc] Assertion errata--items to discuss on 12/17 meeting
From: Bassam Tabbara (bassam@novas.com)
Date: Thu Dec 11 2003 - 09:05:20 PST
Ok Michael,
So let's discuss this next week (thx Swapnajit). Let me just list all of
them here, although I think only one item (before Michael's) I ignored
(cleaning up the callback function, marked with "**" below), and I meant
that it did not affect "voting" and too hard to fix quickly, not to
ignore altogether of course, sorry about that. Anyway, here's the list
of all the items.
1) (Michael): > Just as a side note - and it very likely falls more or
less
> in the hands of the IEEE committee, since this is a problem
> for the whole of PLI - what about using the const qualifier
> for some of the strings (or at least indicating which
> arguments are never touched)? I am pretty sure this is the
> case for several parameters and/or structure elements; so we
> should indicate its intended use
> - including whether it can/will be modified.
Bassam sez: Michael, yes, I understand what you mean, but can't put my
finger on specifics (without a detailed look) since the parameters
(vpiType for example) are already constants, and most of the passed
elements to the calls are pointers allocated user side in most cases.
**2) (Francoise + Michael): > > > For the assertion errata, I think it
is confusing (when I
> read) to
> > > embed the prototype of the callback function in
> > > vpi_register_assertion_cb.
> > >
> > > I would prefer if there would be a function prototype used
and
> > > defined.
Bassam sez: I 100% agree ! I was *also* confused when I was listing the
errata :-(. I'm just worried addressing it may mess things up in the
short time we way, but we can give it a go of course.
-------I think the following are not really debatable...-------
3) (Francoise) > The following sentence :
> 2) a pointer to the time of the callback
> should be changed to:
> 2) a pointer to the time of the assertion attempt
Bassam sez: I would rather we word it like any other "callback", meaning
you get callback reason and the "time the callback got flagged" ....
Would rather not inject assertion attempt time (confusing since there
are many callbacks for different reasons, easier to say "time of the
reason", do you see where I am coming from ?
4) (Francoise)
> Replace (incomplete sentence)
> - For the following operator, the second argument shall be a
> valid assertion handle, the third argument shall
>
> be an attempt start-time (as a pointer to a correctly
> initialized s_vpi_time structure) and the fourth argument
>
> shall be a step control constant.
>
> Usage example: vpi_control(vpiAssertionEnableStep,
> assertionHandle, attemptStartTime,
>
> With:
> - For this operation, the second argument shall be a valid
> assertion handle, the third argument shall
>
> be an attempt start-time (as a pointer to a correctly
> initialized s_vpi_time structure) and the fourth argument
>
> shall be a step control constant.
>
> Usage example: vpi_control(vpiAssertionEnableStep,
>assertionHandle, attemptStartTime, assertionEnableStep)
Bassam sez: I see 2 things in Francoise recommendation:
A) "For this operation": Actually the whole paragraph uses for "For the
following operator" and then talks about it ... That's the flow chosen.
B) "assertionEnableStep": No, it is correct as is (consistent with the
rest), there is a page break, the last argument is on the next page (see
the document) and it should be *vpiAssertionClockSteps* anyway. Just an
artifact, may be I did not explain this well.
Thx.
-Bassam.
-- Dr. Bassam Tabbara Technical Manager, R&D Novas Software, Inc.http://www.novas.com (408) 467-7893
This archive was generated by hypermail 2b28 : Thu Dec 11 2003 - 09:05:59 PST