Re: [vhdl-200x] Call for Vote: VHDL + VHPI (P1076c-2006-D2.4a)]

From: John Shields <John_Shields_at_.....>
Date: Sat Jul 08 2006 - 09:42:41 PDT

Hi Jim,

This was not a private discussion and the VHPI group was aware of it, as was the editor of the LRM.  I raised on the VHPI reflector immediately. As I said, we felt it best to be handled at the IEEE.  THE LRM was in the hands of Accellera's board at that moment.  It was deemed not worth derailing the Accellera board approval of the first draft, making a minor revision, and recycling the draft through Accellera. Procedurally, it cannot be an ISAC issue yet ; of course you know that.  I suppose it could be bugzilla'ed and I simply did not think of that.  This issue came up in April.  As I said, you surprised me with a call for a vote.  I explained the essence of the proposed fix in my earlier mail, but there are details to analyze.  The VHPI group has simply not taken the issue up yet, so there is no complete fix.

I wish we had no LRM bugs, but they happen.  It is straightforward to resolve technically.  This procedural issue of a change to Accellera's draft coming in the IEEE review and approval process is going to be normal and we should expect it.

Next week, I will submit this as a bugzilla.

Regards,
John 

Jim Lewis wrote:

John,
If you have not already submitted this, please submit
it to either ISAC or Bugzilla.   If this is a known bug,
it should already be in the system.  These things cannot
be limited to private discussions.

Was the VHPI group aware of these issues before this?
If not, why not?  Is there a proposed fix?

Regards,
Jim

Hi All,

Vote: negative

Comment:

I am afraid I must vote negative, but there is no other reason than this technical error.  There is a problem that came up very late with vhpi_user.h file.  We declare the abstract type for characters, vphiCharT as char and it must be unsigned char to properly represent the VHDL character set.  A compiler warning:
  vhpi_def.c: In function `vhpi_is_printable':
  vhpi_def.c:19: warning: comparison is always true due to limited range of data type

led to this and we did not deliver this file until quite late. I discussed this with Peter Ashenden when I found it and we agreed that we should let it through Accellera and fix it in the IEEE.  So I did.  You know, at the time, we both thought it would be reviewed in the IEEE before any call for vote, but I appreciate how redundant that might now seem.  I never expected to first raise this issue in a vote. :(

The VHPI group should just be asked to propose the fix, just to make sure it ripples through the API correctly.

Regards,
John Shields

Jim Lewis wrote:

Dear colleagues,
This is a call for vote from IEEE P1076 WG members on the
Accellera approved revision of VHDL that includes VHPI plus
some ISAC revisions.  The purpose of this revision is to make
VHPI available as a standard.  As such, it does not have
the additional revisions that were just completed by the
Accellera VHDL TC.  Those revisions will be put forth later
(Q1 2007?).  This gives industry some time to tune up the
revisions if necessary before they become an IEEE standard.

This revision has been reviewed and approved by both the
Accellera VHDL TC and the Accellera board.  We have a
separate PAR for this work (P1076c).  Currently I am
working on getting the ballot group formed.

Approval in this case shall mean that we accept this revision
to be the revision to send to IEEE for balloting.

The draft is numbered 2.4a by the Accellera VHDL TC and is
available at:
http://www.accellera.org/apps/org/workgroup/vhdl/download.php/488/P1076c-2006-2.4a.zip

Please forward votes to me by email (eg, by replying to this message) by 5pm
US-PDT, Friday July 28, 2006.

Potential votes:  Approve, Negative with comment, Negative with no comment, Abstain

Vote:

Comment:


Best Regards,
Jim Lewis
VASG/ IEEE 1076 WG Chair

Received on Sat Jul 8 09:42:45 2006