[sv-ec] Minutes of sv-ec June 11th, 2007 conference call

From: Neil Korpusik <Neil.Korpusik_at_.....>
Date: Sat Jun 23 2007 - 15:17:23 PDT
-- 
---------------------------------------------------------------------
Neil Korpusik                                     Tel: 408-276-6385
Frontend Technologies (FTAP)                      Fax: 408-276-5092
Sun Microsystems                       email: neil.korpusik@sun.com
---------------------------------------------------------------------


-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.


SV-EC Committee Meeting
Monday Jun 11 2007  11:00am - 1:00pm PST

With the new calculations for voting rights below...
                           3/4 rule  = .75 * 21 = 16
(121202020102020101311) Day
(481593604882505956041)
(000011111100000000000) Month
(889900112211223344456)
(000000000000000000000) Year
(666666666677777777777)
------ Attendees ------
(-AAAAAAAAAAAAAAAAA-AA) Arturo Salz            19
(--AAA-AAAAAAA-AAAAAAA) Cliff Cummings         17
(AAAAAAA-AAAAAAAAAAAAA) Dave Rich              20
(AA-A-AAA-AAAAAAA---AA) Francoise Martinolle   15
(-AAAAAAAAAAAAAAAAAAA-) Mehdi Mohtashemi       19
(AAAAAAAAAAAAAAAAAAAAA) Neil Korpusik          21
(AAAAAAAAAA-AAAAAAAAAA) Ray Ryan               20
(AAAAAAAAAAAA-AAA---AA) Gordon Vreugdenhil     17
(AAAAAA--AAAAA-A--AAAA) Steven Sharp           16
(--AAAA-A-------------) Phil Moorby            05 - No voting rights  
(---AA-AAA-AAAA-AA-A--) Doug Warmke            12 - No voting rights
(AAAAAAA---AA-A-AAAAAA) Stu Sutherland         16
(-AAAA--AAAA-A-AAAAA-A) Heath Chambers         15
(-AAAAAA-A----AAAAAAAA) Don Mills              15
(--AA--A---A-AAA--A-AA) Jonathan Bromley       10
(--A------------------) Logi                   01 - No voting rights  
(----AAA--------------) Melvin                 03 - No voting rights
(-----A-AAAAAA-AAAAAAA) Mark Hartoog           14
(-------A-------------) Satia                  01 - No voting rights  
(--------AAA----------) Rob Slater             03 - No voting rights
(-------------A-------) Alex Gran - Mentor     01 - No voting rights
(---------------A-AAA-) Mike Mintz             04
(------------------AAA) Geoffrey Coram         03  
(-------------------AA) David Scott - Mentor   02 

16 people (other than the chair) currently have voting rights 

 **  Minutes taken by Neil Korpusik 
     Mehdi was not on the call, Neil acted as chair.
//////////////////  June 11, 2007 /////////////////////////

1. IEEE patent policy
   ------------------
    ref:  http://standards.ieee.org/board/pat/pat-slideset.ppt 
       Move: Cliff - Assume that the patent policy was read
     Second: Gord
    Passed unanimously

2. Review meeting minutes/Notes: 
   ----------------------------
   http://www.eda.org/sv-ec/Minutes/SV-EC_Meeting_April_30_2007_Minutes.txt
   http://www.eda.org/sv-ec/Minutes/SV-EC_Meeting_May_14_2007_Minutes.txt
   <not yet posted - were sent out along with the agenda>

       Move: Heath - approve meeting minutes of April 30
     Second: Jonathan
    Abstain: Gord, FM
    Opposed: none
    Passed 

       Move: cliff - approve meeting minutes of May 14, 
	       With the correction to the attendance: Geoffrey Coram was present
     Second:  Stu
    Abstain: Heath
    Opposed: none
    Passed 

3. New operating guidelines - voting rights changes, etc.
   New guidelines were sent out May 25, 2007.
   The main change that pertains to the svec is that there is a new rule 
   in place for retaining voting rights. Anyone that has attended three-fourths
   of all the meetings will have voting rights. 

4. New version of mantis - Dave
   This has been put in place. 
   Relationship field - makes it easier to link related mantis items
   Managers will see the biggest change.
   Searches are now faster.

   Jonathan - The Product Version field (LRM version)
   Dave     - it can be updated. 

5. Discussion on Merged LRM 1800-2008 Draft3    

    Stu     - draft 3a - IEEE version for sale
	    - minor changes sent to Stu that were pure editorial issues were
	      incorporated.
	    - There were quite a few of these (50-60)
	    - He will email a list of the changes that were made. 
	    - Draft 3a was completed (was sent to IEEE and Dennis ~10 days ago)
	    - Jonathan - says it is available on the IEEE web site
	    - purely editorial changes can still go to stu (copy the chair)
    Neil    - all content changes must now go through mantis

6. Review mantis items with proposals      

Those voted on in previous meetings

   1777  changes made properly? (see minutes)
	 http://eda.org/svdb/bug_view_page.php?bug_id=0001777
	 Some people were unable to upload the proposal (blanks in name)

   1787  changes made properly? (see minutes)

AI/Neil - send email on what he thinks is wrong with 1777, 1778

   Mantis items that didn't pass in the email votes

   1371  Semantic of program block $exit
	 http://eda.org/svdb/bug_view_page.php?bug_id=0001371
	 was a new proposal uploaded? (discussed May 14th)
	 Last update was 4/25

AI/Dave - upload a new proposal.

   1384  bit stream cast and pack/unpack for protected./local members
	 http://eda.org/svdb/bug_view_page.php?bug_id=0001384
         was a new proposal uploaded? (discussed May 14th)

AI/Dave - update the section numbers.

   1707  streaming operator should specify which direction slices are taken from
	 http://eda.org/svdb/bug_view_page.php?bug_id=0001707

	 Cliff    - wants to discuss it.
	 Neil     - has a problem with a particular sentence
	 Jonathan - has a big problem with the current description
		  - would like to see a reference algorithm.
		  - he knows enough to take the first cut at such an algorithm.

AI/Jonathan - take a crack at the new algorithm
AI/Neil     - resend my feedback

   1723  Size method for associative arrays  
	 http://eda.org/svdb/bug_view_page.php?bug_id=0001723

	 Cliff   - size - 1024, 16
	 Dave    - high and low index information is available with $high(array)
		 - anything with > or < defined. 
        Jonathan - what if indexed by a string?
	 Dave    - Cliff wants total possible span
        Jonathan - that is a property of the data type not the usage
	 Arturo  - what about string data type? 
	 Cliff   - num - should tell us the number of elements
	 Heath   - other languages (e.g. e) have num the same as size
		   a list in e
         Ray     - for a Q, size tells you the number of elements in the Q
	 Arturo  - would it make sense to have num(?) for all arrays?
	 Ray     - first index is always 0? (answer is yes)
	 Cliff   - bounded Q's are allowed
		 - can we do Cadence damem pli 
		   range of a certain size; use damem pli code to access it.
		   Sparse array.
        Jonathan - an associative array with a subscript of a certain type. 
		   isn't it the same?
         Cliff   - he wasn't sure if implementations allow what he wants
	 Steven  - can't handle a range that isn't a power of 2. 
		 - need to always start at 0.
         Dave    - num and size - he thinks are interchangeable.
		 - limit or something else would be more descriptive
		 - can't determine the limit of a mailbox
		 - num and size are just arbitrary names
		 - wants the same word to find the number of elements.
		   i.e. a consistent method name
                 - need both now, for backwards compatibility
	 Cliff   - thinks of them as being different
		   What about the following
		      num  - dynamic array and Q
		      size - can't be used for assoc.

    Move: Dave - approve the proposal for 1723, with friendly amendment 
		 for section number 7.10.1
  Second: Heath
 Abstain:  
 Opposed: Cliff - doesn't think of size as being the number of elements, 
		  thinks of it as the address spanned, 
		  encompassed by elements already allocated.
	  Stu   - no change needed
 Passed - with 2 no votes

AI/Dave - upload a proposal with the updated section number.

   1500  Forward typedef of a class is only to declare instances of that type 
	 http://eda.org/svdb/bug_view_page.php?bug_id=0001500

          Gord   - wants this mantis item to pass 
		 - Knows of at least one big user that has a flow for their 
		   design style: global type parameters - using forward typedefs
		   This change would affect that style of flow.
		 - compilation units - will resolve a typedef at end
          Arturo - agrees with Gordon
		   thinks a forward typedef is only useful for classes
          Ray    - the LRM has the following example 
		     typedef foo;
		     foo f = 1;
		     typedef int foo;
          Mark   - typedefs existed in SystemVerilog before classes existed
	  Stu    - there are backward compatibility issues.
	  Gord   - would like to be allowed to make changes if it makes sense
	  Dave   - Superlog had pointers to structures. That is why this 
		   capability existed in SystemVerilog 3.0
          Mark   - what is the compelling reason for taking it away?
          Gord   - an implementation has to not do anything until $unit is 
		   loaded - then you can make decisions 
                 - forward typedef - inefficiency issue
	           users have a sense that they aren't resolved after elab.
	  Arturo - put it into the LRM that they are similar to parameters. 
		   e.g. inefficient
          Gordon - if this came up 2 years ago he would have pushed for it.
		 - not sure how it would be enforced
		 - users are expecting typedefs to be handled during compilation
		   and not treated like a type parameter (e.g. inefficient)
                 - name resolution committee - binding issues still remain
		 - how do name resolution into a forward typedef if it has 
		   the same restrictions as a type parameter. 
          Mark   - agrees that a forward typedef is similar to a parameter
	  Gordon - arguing that we should pass this with the thought that 
		   it is needed for some other name resolution issues. 
          Mark   - is working on alternate language rules for type parameters
	  Gordon - we should then table this item.
	  Ray    - 16.8 has a number of other places that would also need to 
		   be made if we decide to approve mantis 1500.
          <we agreed to wait for Mark's write-up on type parameters>

AI/Mark - send out to the sv-ec his write-up on type parameters.  

   1556  in-line static variable initialization - require keyword static?
	 http://eda.org/svdb/bug_view_page.php?bug_id=0001556

         module top;
	 int count = -1;                   // static keyword not required
	 function int counter();
	   return count++;
	 endfunction
	 initial begin
	   for (int i=0; i<3; i++) begin
	     automatic int loop1 = counter();
	     print();
	   end                             // prints 1 2 3
	   for (int i=0; i<3; i++) begin
	     int loop2;                    // static keyword not required
	     static int loop3 = counter(); // static keyword required
	     print();
	   end                             // prints 3 3 3 
	   for (int i=0; i<3; i++) begin
	     int loop3 = counter();        // illegal
					   // static or automatic to be legal
	     print();
	   end
	 end

	 Cliff   - wasn't sure that adding the keyword was needed
	 Dave    - was in original sv 3.0 and removed since it wasn't clear
        Jonathan - fixing a point of misunderstanding. Thinks that there
		   are issues with other variables. 
		 - the loop2 variable in the example is also static
		 - thinks it is crazy to have two statics next to each other. 
		   and to not require static on both.
                 - storing up a bunch of surprises in the future. 
         Gord    - the issue is when the initialization happens.
		   agrees this is irregular, but less surprising. 
                 - would like to require static on both, but can't due to 
		   backward compatibility
         Steve   - was in 3.0, removed by svec?, he thinks so. Thought it 
		   should have been done in svbc. Procedural problem.
        Jonathan - thinks it should be deprecated 
	 Stu     - thinks it should be a lint check - init of a static
	 Steven  - C requires static for both of them.
        Jonathan - doesn't use a lint checker. 
	 Dave    - can't require static for all cases (backward compat)
	 stu     - not likely for this one change to cause backward compat
        Jonathan - it would be legal to add static on loop2? (yes)

    Move: Dave - approve motion for 1556, new section number 6.6 --> 6.21
  Second:  stu
 Abstain: 
 Opposed:  
 Passed unanimously


   1608  equality, inequality and conditional operator rules for class handles
	 http://eda.org/svdb/bug_view_page.php?bug_id=0001608
   1609  import statements should not be allowed in class scopes
	 http://eda.org/svdb/bug_view_page.php?bug_id=0001609
   1715  Triggered property of a clocking block
	 http://eda.org/svdb/bug_view_page.php?bug_id=0001715

   1789 is ready for an email vote
   1857 is ready for an email vote
   1371 is ready for an email vote

AI/Neil - start an email vote for these 3 mantis items

7. Discussion: mantis items with no proposal
8. Next meeting: 
   June 25th 2007  - yes
   July  9th 2007  (proposed)
Received on Sat Jun 23 15:17:47 2007

This archive was generated by hypermail 2.1.8 : Sat Jun 23 2007 - 15:18:05 PDT