VHDL Issue Number:      0056
Classification:         Language Definition Problem
Language Version:       VHDL-87
Summary:                Unclear requirements on prefix in predefined attribute
                        names.
Related Issues:     
Relevant LRM Sections:  6.1, 6.5, 14.1
Key Words and Phrases:  Attribute name, appropriate for, prefix
Current Status:         Analyzed
1076-1993 Disposition:  Superseded (ISAC Issues Outstanding)
Disposition Rationale:  ??
Superseded By:          1005
-----------------------
Date Submitted:         1989/02/10
Author of Submission:   Doug Dunlop
Author's Affiliation:   Intermetrics, Inc.
Author's Post Address:  4733 Bethesda Ave #415
            Bethesda, MD  20814
Author's Phone Number:  (301) 657-3775
Author's Fax Number:    
Author's Net Address:   dunlop@inmet.inmet.com
-----------------------
Date Analyzed:          1991/11/19
Author of Analysis:     Stephen Bailey
Revision Number:        $Revision: 1.9 $
Date Last Revised:      $Date: 1995/05/13 21:53:48 $


Description of Problem
----------------------
The LRM says that the prefix in a 'RANGE attribute name
may be "appropriate for an array object, or alias thereof".  However, the
notion of appropriateness of prefixes (LRM Section 6.1) is defined for types
and not objects.  Is the intent here to require the prefix to be an object or
to require the prefix to be appropriate for an array type (as in Ada)?


Proposed Resolution
-------------------
The wording appears to have been deliberately altered when taken
from the Ada LRM.  Hence we assume the prefix must be an array object, an
alias thereof, or a constrained array subtype.  The same is true for all
predefined attributes whose prefix is required to be "appropriate for an array
object, or alias thereof, ...".


VASG-ISAC Analysis & Rationale
------------------------------
   The usage of the phrases "appropriate for an array object" (LRM 14.1) and
   "appropriate for a one-dimensional array object" (LRM 6.5) appear to be
   errors in the LRM and not deliberate alterations from the Ada LRM as
   suggested.  The rationale for this position is that:

    1.  The LRM contains no definition of appropriateness as applied
        to objects, but does define it for types, and

    2.  The definition of appropriateness for types as applied to the
        context of prefixes (which is the context of LRM 14.1) includes
        prefixes which are objects.  "A prefix is said to be _appropriate_
        for a type in either of the following cases:  The type of the
        prefix is the type considered.  The type of the prefix is an
        access type whose designated type is the type considered."
        In both cases, it is not the prefix that is being considered but
        the type (or type accessed) of the prefix.
        The definition does not refer directly to the prefix.

   Furthermore, the definition of appropriateness for types allows the
   application of an attribute to a prefix that is an object that is
   of an access type without requiring the use of ".all."  It is difficult
   to conceive of a definition of appropriateness as applied to objects
   that would have the same power of the definition of appropriateness
   of types while also being meaningfully different from the type
   appropriateness definition.

   Therefore, it is reasonable to conclude that inattention to detail in
   the writing of the LRM resulted in the two phrases above in sections 6.5
   and 14.1.


VASG-ISAC Recommendation for IEEE Std 1076-1987
-----------------------------------------------
   Recommend that the 1st sentence of the 2nd paragraph of LRM section 6.5
   be changed to:

    The prefix of a slice must be appropriate for a one-
    dimensional array type.

   Recommend that all occurrences of the phrase "that is appropriate for an
   array object, or an alias thereof, or that denotes a constrained array
subtype"
   in section 14.1 be replaced by:

    that is appropriate for an array type or that denotes a constrained
    array subtype

   NOTE: Since aliases are typed, the phrase "appropriate for an array type"
applies
   to aliases as well as to other objects and "denotes a constrained array
subtype"
   allows the attribute to be applied directly to a constrained subtype name.


VASG-ISAC Recommendation for Future Revisions
---------------------------------------------
   Future versions of the LRM should be modified according to the
   above recommendation.