From bobduff  Wed Dec 16 15:53:18 1992
From: bobduff (Bob Duff)
To: ada9x-mrt
To: iso@ajpo.sei.cmu.edu
Subject:  LSN-1063 on Ada 9X Preference Rule

!topic LSN on Ada 9X Preference Rule
!key LSN-1063 on Ada 9X Preference Rule
!reference MS-8.7;4.6
!reference RM83-4.6(15)
!from Bob Duff $Date: 92/12/16 15:49:29 $ $Revision: 1.4 $
!discussion

This Language Study Note is an update to LSN-1037.
It addresses resolution number 9-8, "Preference Rule",
from the Salem WG9 meeting.

As Tucker explained at the Salem meeting, the new preference rule is:

  There is a preference for the primitive operators of the root numeric
  types root_integer and root_real.  The rule is as follows:

    - In a given complete context, if two acceptable interpretations
      differ only in that one is for a primitive operator of the type
      root_integer or root_real, and the other is not, the interpretation
      using the primitive operator of the root numeric type is
      preferred.

    - For an innermost complete context, if there is exactly one overall
      acceptable interpretation where each constituent's interpretation
      is the same as or preferred (in the above sense) over those in all
      other overall acceptable interpretations, then that one overall
      acceptable interpretation is chosen.  If the ambiguity cannot be
      resolved in this manner, the complete context is illegal.

This is the same as stated in LSN-1037 and in the ILS, except that we
have replaced "for a corresponding primitive operator for another
numeric type" with "not".  The reason for this change is upward
compatibility.

This rule is upward compatible except in cases involving Beaujolais
effects.  Ada 83 programs containing Beaujolais effects will be
ambiguous, and therefore illegal, in Ada 9X.  Thus, the rule is
upward consistent.  [Note that LSN-1037 incorrectly made this same
claim for the rule stated there.  We had modified the LPT rule to
eliminate what we believed were confusing overload resolution rules,
thus breaking the property as proven by the LPT, a fact that we
missed when writing LSN-1037.  The change to "not" restores the
property.]

WG9 requested study of this issue by the Language Precision Team (LPT).
However, the LPT has already studied this issue in depth.  For a
detailed analysis of the elimination of Beaujolais effects, and the
upward consistency of the proposed preference rule, see Chapter 6,
"Overloading Preference Rules in Ada 9X", of the "Technical Operating
Report (TOR) -- Formal Definition Report" of November 12, 1991, from the
LPT.  There are no plans for further study of this issue by the LPT.

In any case, the alternative of reverting to the Ada 83 rules is not
possible, for the reasons explained in LSN-1037.

------ Output from automsg.perl ------
Mail received from bobduff
----- Ignored header line(s): -----
!reference RM83-4.6(15)
-----------------------------------
*** Key LSN-1063 given to topic 'LSN on Ada 9X Preference Rule'
----------- End of output ------------