ISAC: Minute of ISAC meeting on 16 March 2006

Attached are the minutes, which are also available at the website.

Also,  note that there is some discussion about the time of the next 
ISAC meeting.

Daylight savings time takes effect, so we normally move the meeting time 
to 7 pm Pacific
Daylight Time. However, we anticipate that the Accellera VHDL meeting on 
the same day,
will continue to be held from 3pm to 5pm Pacific time. It might be 
better to meet at
6pm instead of 7pm to avoid a two hour gap. Let me know what you want.

Chuck Swart

Minutes of ISAC meeting held via telecom on 16 March 2006

Present: Peter Ashenden, Larry Soule, Chuck Swart, Lance Thompson

Absent: Jim Lewis, Ajay Verikat

Next Meeting: Thursday April 06, 2006, 7 pm Pacific Daylight Time
              (Friday April 07, 2006, 2 am GMT)

Note: This time is flexible. Because of VHDL-Accellera meetings which
occur before it, we may prefer to meet at 6 pm Pacific Daylight Time
instead. Let me know your preferences.

TOPIC: IR 2077 Incorrect wording on some language constructs

This has been ISAC-approved and, barring any last-minutes objections,
will be forwarded to the VASG.

ACTION: Chuck to issue "last call" then forward to VASG.

TOPIC: IR 2084  A record "element" is not called a "field"

This has been ISAC-approved and, barring any last-minutes objections,
will be forwarded to the VASG.

ACTION: Chuck to issue "last call" then forward to VASG.

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

It was argued that the LRM is not clear whether or not section
12.3.1.4 was intended to apply to interface_declarations. Peter
pointed out (again) that note 1 mentions ports, which are interface
objects, so the intent is to include them.

There was some discussion about the most useful initialization rule,
specifically, whether the initial value should be the value of the
actual parameter or whether it should be the default value. It was
observed that one can always get the "copy-in" "copy_out" semantics by
declaring the mode to be INOUT, so it is probably better to initialize
the formal to the default value.

ACTION: Chuck to update. 

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

Peter claims that this issue has been addressed by LCS 113.

ACTION: Chuck to distribute LCS 113. All to review the LCS. 

TOPIC: IR 2062 Range staticness

Chuck explained the rationale for the broader definitions of global
staticness which could be applied to many predefined attributes. Peter
suggested that an alternate name for "appropriate for a globally
static expression" could be "appropriate for a globally static
attribute name".

ACTION: Chuck to update. All to review.

TOPIC: IR 2074 Problem with direct/select visibility in formal part 

It appears that several suggested changes described in minutes of
January 12 2006 were not incorporated.

ACTION: Chuck to update.

TOPIC: IR 2079 Is TIME a locally static type?

This is ready for an ISAC vote.

ACTION: Chuck to submit for vote.

TOPIC: IR 2083 Generate index specification should be of same subtype
as generate parameter

Peter suggested a slight wording improvement: "must lie within the
discrete range" should be "must belong to the discrete range."

ACTION: Chuck to revise then submit for vote.

TOPIC: IR 2087 Ambiguous rule for type of an alias declaration

The ISAC recommendation needs to be revised to include removing the
unnecessary sentence described in the analysis. With this revision,
the IR is ISAC approved.

ACTION: Chuck to update, distribute, issue "last call" then forward to VASG.

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

Lance pointed out that there is a missing case, when the guarded
signal is an index of a declared signal.  Also, the paragraph which
starts "If the guarded signal specification denotes a slice of an
array subelement." refers to the wrong type.

ACTION: Chuck to revise.