ISAC: Proposed agenda for ISAC meeting January 12, 2006

Next Meeting: Thursday January 12, 2006, 6 pm Pacific Standard Time
              (Friday January 13, 2006, 2 am GMT)


telecom info
US 800-637-5822
Other +1 647-723-3937
Access Code 6850821

1. New ISAC Issues:

2080      Case expression should include parenthesized expression

2. Review comments on disputed ISAC Issue:

2074 Problem with direct/select visibility in formal part

(See comments below)

3. Open ISAC issues:

2038    Submitted       Ajay            Minor semantic errors
2054    Analyzed        Larry           Individual assoc. rules for 
array formal are not valid
2075    Submitted       Ajay            Arrays with numeric and 
enumeration index types are not closely related
2077    Submitted       Larry           Incorrect wording on some 
language constructs
2079    Submitted       Chuck           Is TIME a locally static type?

4. Categorization of Old IRs:


Comments from Jim Lewis on IR2074:
2074, Abstain with Comments
I am a little conflicted with this.
I agree that we should at a minimum do what 2074 proposes.
I would prefer that both examples A and B were illegal.
If I were reviewing someone's code, this would be a
terrible thing to read I would prefer they wrote:

    instance1: c1 PORT MAP ( X01(work.p2.X01)=> s2);

I don't suppose this is a good reason for making B illegal though.

With the proposed change:
  "At a place in which a declaration is visible by selection every
  declaration with the same designator which would otherwise be
  directly visible is hidden."

With B, if I understand it right, the second reference to X01 is
work.p2.X01 since the formal port is not visible by selection anymore.
    instance1: c1 PORT MAP ( X01(X01)=> s2);

Also toward the end of the analysis there is the statement:
  In conclusion, the ISAC recommends that ...
  with a similar interpretation for actual_part.

Are there similar problems for the actual_part?  In the actual
part, X01 of the component is not visible on only things
that are directly visible are visible, right?


Comments from Peter Ashenden:

Folks,

I generally agree that visibility by selection should have precedence over
direct visibility, and so agree with the ultimate recommendation in the IR.
However, I think there is a problem in the analysis related to
interpretation of names in the examples. Specifically, a port name is
visible by selection "at the place of the formal designator" of an
association element. I would interpret this as meaning throughout the formal
designator. In the examples, the port designator takes the form of an
indexed name, X01(X01), in which the prefix and the index expression are
both the name X01. Hence, both occurrences of X01 would refer to the port
name. That means the formal designator is not a locally static name, and for
that reason, the example is illegal.

Cheers,

PA