RE: [sv-cc] mantis 2226 action completed

From: Shields, John <John_Shields_at_.....>
Date: Mon Sep 29 2008 - 15:25:36 PDT
Hi,

 

I've incorporated all of Jim's changes in this email except item (5),
which I have left as is.  I have addressed all his feedback from the
previous 3 emails as well. It is easier for me to manage the change in
real time and I can undo any of this if needed, but it is very
reasonable and friendly.  I am not going to repost anything just yet,
but I certainly can on short notice.  My plan is to wait for any other
feedback and to allow Charlie to consider whether we need discussion.

 

Regards, John

 

________________________________

From: owner-sv-cc@server.eda.org [mailto:owner-sv-cc@server.eda.org] On
Behalf Of Shields, John
Sent: Monday, September 29, 2008 3:00 PM
To: Jim Vellenga
Cc: sv-cc@server.eda.org
Subject: RE: [sv-cc] mantis 2226 action completed

 

Thanks, Jim. Comments below.

 

________________________________

From: Jim Vellenga [mailto:vellenga@cadence.com] 
Sent: Monday, September 29, 2008 1:51 PM
To: Shields, John
Cc: SV-CC
Subject: RE: [sv-cc] mantis 2226 action completed

 

John, you have completed a heroic effort.  Thanks for all the work 

you've put in.

 

I finally decided to start batching myadditional observations,

thoughts, and questions.  Some of these will be trivial, others

we may need to talk about as a committee. (These are only for

parts 2 and 3.)

 

(1) In "37.28 Class variables and class objects", in detail 3, why was 

the sentence "If the class var has the value of NULL, the vpiClassObj 

relationship applied to the class var shall return a null handle" 

removed?

 

It was removed long ago in the approved proposal. I'll guess that it was
somehow viewed to be obvious and did not need to be stated, but I don't
recall.  It is fine with me to keep it.

I'll add it as a friendly amendment.

 

(2) In "37.29 Constraint, constraint ordering, distribution', why isn't 

the sentence "For details on memory allocation property, see <cross 

reference to 37.3.7>" an additional detail, as many similar sentences 

are?

 

Should be.  It was an omission in the approved proposal.  I'll add it as
a friendly amendment.

 

(3) In "38.x vpi_get64()", there's a double period after PLI_INT64.

Got it.

 

(4) In "38.36.1 Simulation event callbacks," there's an extra period 

before "When" in the description for "cbReclaimObj".

Got it.

(5) In "38.36.1 Simulation event callbacks," now that you've added a 

sentence to the description of "cbReclaimObj," I'm wondering if we 

should clarify it by saying "any associated callbacks should be 

considered invalid, except that any associated pending cbEndOfObject 

callbacks shall first be executed."

 

I didn't add anything.  The approved proposal had this wording.  I don't
think we need to wordsmith this. There is a reality when there are
multiple callbacks registered for the same purpose, perhaps even from
different and unrelated applications, that they will all occur and in no
defined order. All the end of life callbacks for a particular object are
in this bucket. Your clarifying comment really refers that situation.  I
don't think it needs to be said and there certainly can be no ordering
implied.  The associated callbacks, as I intended it, was for things
like value change callbacks.  We happened to remove that particular one
for classes very late in the discussion, because you had not had time to
give it proper consideration.  That is unfortunate because it was pretty
key to logging classes, but you made a good point.  It think the
associated callbacks is an intent that we should not lose.  If anything,
we could add it for other end of life callbacks.  I am comfortable
taking that all up in the future.

 

(6) In "38.36.1 Simulation event callbacks," in the description for 

cbEndOfThread, there's a stray space between "callback" and the ensuing 

comma.

Got it.

(7) In "38.36.1 Simulation event callbacks," in the paragraph beginning 

with "The cbValueChange callback may be placed ...," the string 

"s_cb_data" should be rendered in a fixed width font, while "vpiObjId" 

should be rendered in bold.

Got it.

(8) In "38.36.1 Simulation event callbacks," in the paragraph beginning 

with "For a cbReclaimObj callback...," the string "s_cb_data" should be 

rendered in a fixed width font, and the word "vpi" in "All vpi 

properties" should be changed to all upper case.

Got it.

(9) Now that I'm taking a fresh look at "37.x vpi_release_handle()", 

I'm wondering if it would be better to say "The VPI routine 

vpi_release_handle() shall free memory allocated for VPI handles" 

instead of "for objects."  I realize we passed it as is once before, 

but I don't want the standard to give the impression that executing 

this routine will affect the underlying simulation in any way.

 

(10) Also in "37.x vpi_release_handle()," would it be better to replace 

"vpi_release_handle() shall generally be used to free memory created 

for iterator objects" with "vpi_release_handle may be used to free 

memory created for iterator objects"?  The existing wording, taken from 

the vpi_free_object() description, has bothered me for a long time 

because "shall" implies a requirement that doesn't really exist.

I am fine with both (9) and (10) as friendly amendments.

 

Regards, 

Jim Vellenga

 

---------------------------------------------------------
James H. Vellenga                            978-262-6381
Software Architect                              (FAX) 978-262-6636
Cadence Design Systems, Inc.         vellenga@cadence.com
270 Billerica Rd
Chelmsford, MA 01824-4179
"We all work with partial information."
---------------------------------------------------------- 

 

	 

	
________________________________


	From: owner-sv-cc@eda.org [mailto:owner-sv-cc@eda.org] On Behalf
Of Shields, John
	Sent: Monday, September 29, 2008 2:07 PM
	To: sv-cc@eda.org
	Cc: sv-champions@eda.org; Neil.Korpusik@sun.com; Karen Pieper;
Bresticker, Shalom; stuart@sutherland-hdl.com
	Subject: [sv-cc] mantis 2226 action completed

	Hi Everyone,

	 

	I have completed my action on 2226.  I changed it to review so
that I could revise files.  I've removed all the draft6 revisions and
added the changes baselined to draft7.   All the champion and reviewer
feedback has been incorporated. The Champions and the Working Group have
entrusted SV-CC to to this composition work and you have entrusted that
all to me.  In a perfect world, I would ask Charlie to close 2226 and he
in turn would ask Neil to remark it approved and turn it over to Stu.  I
think it would be helpful to have a quick review for composition errors
and to allow the friendly amendments to be seen in context.  I leave the
next action to Charlie to ask for review or close the item.

	 

	Because of the number of files, I have packaged them in 3 zip
files.  fm-clauses-mantis2226.zip contain full framemaker files for each
clause (except annex C).  This is what Stu asked for to minimize his
effort.  My notes on that are:

	 

	1) Some diagrams I could show deletions, others the real estate
for the deletion was reused and one sees only the addition.

	2) The annexes need normal editor work.  I did not have Annex C
from Stu at all, but the edit is trivial.  The other 2 annexes require
some minor numbering.

	 

	 

	fm-changes-only-mantis2226.zip contains framemaker files
representing only what has changed in a minimal fashion suitable for
efficient review. pdf-changes-only-mantis2226.zip contains pdfs of
those.  The framemaker change only versions were necessary(to me) for
composition and I used them to create the complete clauses.  Normally,
Stu would start with them.  It is his option to look at them and
validate that the edits to the full clauses were made accurately.  I
used paragraph tags properly in the full clauses, but the change-only
requires tag mods to insure that the numbering matched the document.

	
	Regards, John

	
	-- 
	This message has been scanned for viruses and 
	dangerous content by MailScanner <http://www.mailscanner.info/>
, and is 
	believed to be clean. 


-- 
This message has been scanned for viruses and 
dangerous content by MailScanner <http://www.mailscanner.info/> , and is

believed to be clean. 

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
Received on Mon Sep 29 15:28:20 2008

This archive was generated by hypermail 2.1.8 : Mon Sep 29 2008 - 15:28:35 PDT