ISAC: Minutes of ISAC meeting held on 06 April 2006

These are also available online.

Chuck Swart


Minutes of ISAC meeting held via telecom on 06 April 2006

Present: Peter Ashenden, Jim Lewis, Larry Soule, Chuck Swart, Lance
Thompson, Ajay Verikat

Absent: (all present)

Next Meeting: Thursday April 20, 2006, 7 pm Pacific Daylight Time
              (Friday April 21, 2006, 2 am GMT)

TOPIC: IR 2088 Formatted I/O

This is clearly an enhancement request and will be forwarded through
the VASG to Accellera. Jim was able to find a copy of the C-Library
SNUG paper referenced by the submitter.

ACTION: Chuck to forward to VASG.

TOPIC: IR 2089 Directional records

This is clearly an enhancement request, and follows the flavor of
interfaces which are currently being developed.

ACTION: Chuck to forward to VASG. Lance to inform the author of this
action.

TOPIC: IR 2090 Signature in alias declaration for "not" wrong

The submitter is correct that the alias he mentions is
wrong. The correction was  ISAC approved.

ACTION: Chuck to update. 

TOPIC: IR 2075 Arrays with numeric and enumeration index types are not
closely related

This was approved by the VASG with two negative votes. The clear
consensus was to approve. Also, it was felt that extending the
capabilities of type conversions in this way did not break strong
typing.

ACTION: the IR is VASG-Approved. Chuck will submit an analysis for the
VASG chair to submit to those who voted on the three IRs 2075, 2080
and 2081.

TOPIC: IR 2080 Case expression should include parenthesized expression

This was approved in the VASG vote without comment.

ACTION: Chuck to update.

TOPIC: IR 2081 The term ancestor is used where parent is meant

This was approved by the VASG without comment.

ACTION: Chuck to update.

TOPIC: Type conversions

Jim Lewis reported that Dave Bishop found that implementations
performed differently on type conversions between std_ulogic_vectors,
std_logic_vectors and unsigned vectors. Some implementations allowed
all such conversions while others only allowed conversions in which the
element subtypes are the same (both std_logic or both std_ulogic).  An
IR will be issued to resolve this issue.

ACTION: Chuck to work with Dave Bishop to get the IR.

TOPIC: IR 2038 Minor semantic errors

No progress. Ajay will analyze by next meeting.

ACTION: Ajay to analyze.

TOPIC: IR 2054 Individual assoc. rules for array formal are not valid

Peter claims that LCS 113 covers this.

ACTION: All to read LCS 113. We'll discuss this at next meeting.

TOPIC: IR 2062 Range staticness

John Ries commented that the following case was not covered.

type my_record is record
  f1: std_logic_vector(g1 downto g2);
end record;

signal s1 : my_record;

...  s1.f1'left /* this should be globally static but I don't think
the current changes catch it */

Peter remarked that since s1.f1 is a signal name, this case is
covered.  There was also a discussion about the general confusion
between signals and ports, since all ports are signals. Peter pointed
out that the terms "explicitly declared signal" and (rarely)
"implicitly declared signal"are intended to cover the cases in which
the LRM means "signal but not port"

ACTION: All to review this IR. We'll discuss it at the next meeting.

TOPIC: IR 2074

No progress.

ACTION: Chuck to analyze.


TOPIC: IR 2082 Elaboration of unconstrained interface objects

No progress.

ACTION: Chuck to analyze.

TOPIC: IR 2085 What happens when a parameter of mode out is not
assigned in a procedure?

No progress.

ACTION: Chuck to analyze.

TOPIC: IR 2086 Incorrect description of type mark in disconnection
specification

Chuck reported some confusion about why there were special cases for
the "top" signal and for subelements. Peter replied that in one
instance the type in the signal declaration is used; in the other the
type declaration is used.

ACTION: Chuck to analyze.

TOPIC: Old IRs.

Because of other activities, little progress has been made in
categorizing these IRs. There are some significant issues outstanding.

ACTION: Chuck to provide summaries of actions taken so far.